Locating License Plate Based on Edge Features of Intensity and Saturation Subimages

This paper proposes a novel license plate localization algorithm for automatic license plate recognition (LPR) systems. The proposed approach uses color edge information to refine the edge points extracted in a gray-level image. In the proposed LPR system, input images are processed as follows: (1) de- interlacing by the proposed vertical median filter firstly; (2) locating license plates by detecting edge in the intensity and the saturation subimages; (3) identifying the numbers by matching the seven invariants of characters. In our experiments, 368 images are tested during two hours test. The license plate location rate of success is 97.0% with 11 images failed to locate.